NVIDIA / GenerativeAIExamples

Generative AI reference workflows optimized for accelerated infrastructure and microservice architecture.
Apache License 2.0
2.45k stars 527 forks source link

Customer Issue Fix #218

Closed vinay-raman closed 1 month ago

vinay-raman commented 1 month ago

NeMo retriever synthetic data generation:

Customer had a following feedback: Documents that have preset ids need to have their original ids in the generated results.

Fix: The rawdoc format accepts "_id" as a key which can include preset document ids. These ids are persisted in the generated results.